Aluminium is presently produced by electrolysis of an aluminium-containing compound dissolved in a molten electrolyte, and the electrowinning process is performed in cells of conventional Hall-Hèroult design. These electrolysis cells are equipped with horizontally aligned electrodes, where the electrically conductive anodes and cathodes of today's cells are made from carbon materials. The electrolyte is based on a mixture of sodium fluoride and aluminium fluoride, with smaller additions of alkaline and alkaline earth fluorides. The electrowinning process takes place as the current passed through the electrolyte from the anode to the cathode causes the electrical discharge of aluminium-containing ions at the cathode, producing molten aluminium, and the formation of carbon dioxide at the anode (see Haupin and Kvande, 2000). The overall reaction of the process can be illustrated by the equation:2Al2O3+3C=4Al+3CO2  (1)
Due to the horizontal electrode configuration, the preferred electrolyte composition and the use of consumable carbon anodes, the currently used Hall-Hèroult process displays several shortcomings and weaknesses. These weaknesses include area-intensive design, high investment costs, troublesome electrolyte and metal flow patterns, expensive electric busbar systems, etc.
The traditional aluminium production cells utilise carbon materials as the electrically conductive cathode. Since carbon is not wetted by molten aluminium, it is necessary to maintain a deep pool of molten aluminium metal above the carbon cathode, and it is in fact the surface of the aluminium pool that is the “true” cathode in the present cells. A major drawback of this metal pool is that the high amperage of modern cells (>150 kA) creates considerable magnetic forces, disturbing. As a result, the metal tends to move around in the cell causing wave movements that might locally shortcut the cell and promote dissolution of the produced aluminium into the electrolyte. In order to overcome this problem, complex busbar systems are designed to compensate for the magnetic forces and to keep the metal pool as stable and flat as possible. The complex busbar system is costly, and if the disturbance of the metal pool is too large, aluminium dissolution in the electrolyte will be enhanced, resulting in reduced current efficiency due to the back reaction:2Al+3CO2=Al2O3+3CO  (2)
The preferred carbon anodes of today's cells are consumed in the process according to reaction (1), with a typical gross anode consumption of 500 to 550 kg of carbon per tonne of aluminium produced. The use of carbon anodes results in the production of pollutant greenhouse gases like CO2 and CO in addition to the so-called PFC gases (CF4, C2F6, etc.) which are even more pollutant greenhouse gases and very stable. The consumption of the anode in the process means that the interpolar distance in the cell will constantly change, and the position of the anodes must be frequently adjusted to keep the optimum operating interpolar distance. Additionally, each anode is replaced with a new anode at regular intervals. Even though the carbon material and the manufacture of the anodes are relatively inexpensive, the handling of the used anodes (butts) makes up a major portion of the operating cost in a modern primary aluminium smelter.
The raw material used in the Hall-Hèroult cells is aluminium oxide, also called alumina. Alumina has a relatively low solubility in most electrolytes. In order to achieve sufficient alumina solubility, the temperature of the molten electrolyte in the electrowinning cell must be kept high. Today, normal operating temperatures for Hall-Hèroult cells are in the range 940-970° C. To maintain the high operating temperatures, a considerable amount of heat must be generated in the cell, and the major portion of the heat generation takes place in the interpolar space between the electrodes. Due to the high electrolyte temperature, the side walls of today's aluminium production cells are not resistant to the combination of oxidising gases and cryolite-based melts, so the cell side linings must be protected during cell operation. This is normally achieved by the formation of a crust of frozen bath ledge on the side walls. The maintenance of this ledge necessitates operating conditions where high heat losses through the side walls is a cardinal requirement. This results in the electrolytic production having an energy consumption that is substantially higher than the theoretical minimum for aluminium production. The high resistance of the bath in the interpolar space accounts for 35-45% of the voltage losses in the cell. The state-of-the-art of present technology is cells operating at current load sin the range 250-350 kA, with energy consumption around 13 kWh/kg Al and a current efficiency of 94-95%.
As pointed out, there are several good reasons for improving the cell design and the electrode materials in aluminium electrolysis cells, and several attempts have been made to obtain these improvements.
With an inert anode in the electrowinning of aluminium, the overall reaction would be:2Al2O3=2Al+3O2  (3)
Many attempts have been made to find the optimum inert anode material and the Introduction of these materials in electrolytic cells, and numerous patents have been proposed for inert anode materials for aluminium electrowinning. Most of the proposed inert anode materials have been based on tin oxide and nickel ferrites, where the anodes may be a pure oxide material or a cermet type material. The first work on inert anodes was initiated by C. M. Hall, who worked with copper metal (Cu) as a possible anode material in his electrolysis cells. Generally, the inert anodes can be divided into metal anodes, oxide-based ceramic anodes and cermets based on a combination of metals and oxide ceramics. The proposed oxide-containing inert anodes may be based on one or more metal oxides, wherein the oxides may have different functions, as for instance chemical “inertness” towards cryolite-based melts and high electrical conductivity. The proposed differential behaviour of the oxides in the harsh environment of the electrolysis cell is, however, questionable. The metal phase in the cermet anodes may likewise be a single metal or a combination of several metals (metal alloys). The main problem with all of the suggested anode materials is their chemical resistance to the highly corrosive environment due to the evolution of pure oxygen gas (1 bar) and the cryolite-based electrolyte. To reduce the problems of anode dissolution into the electrolyte, additions of anode material components (U.S. Pat. No. 4,504,369) and a self generating/repairing mixture of cerium based oxyfluoride compounds (U.S. Pat. Nos. 4,614,569, 4,680,049 and 4,683,037) have been suggested as possible inhibitors of the electrochemical corrosion of the inert anodes. However, none of these systems have been demonstrated as viable solutions.
The introduction of inert anodes and wettable cathodes in the present Hall-Hèroult electrowinning cells would have a significant impact on reducing the production of greenhouse gases like CO2, CO and PFC's from aluminium production. Also, potentially the reduction in energy added can be substantial if the inter-electrode space can be reduced in comparison to traditional Hall-Hèroult cells.
Patents regarding retrofit or enhanced development of Hall-Hèroult cells are amongst others described in U.S. Pat. Nos. 4,504,366, 4,596,637, 4,614,569, 4,737,247, 5,019,225, 5,279,715, 5,286,359 and 5,415,742, as well as GB 2 076 021. All of these patents address the problems encountered due to the high heat losses in the present Hall-Hèroult cells, and the electrolysis process is operated at reduced interpolar distances. Some of the proposed designs are in addition effective with respect to reducing the surface area of the liquid aluminium metal pad exposed to the electrolyte. However, only a few of the suggested designs have addressed the low production to area ratio of the Hall-Hèroult cells. Amongst others, U.S. Pat. Nos. 4,504,366, 5,279,715 and 5,415,742 have tried to solve this problem by implementation of vertical electrode configurations to increase the total electrode area of the cell. These three patents have also suggested the use of bipolar electrodes. The major problem of the cell design suggested in these patents, however, is that the requirement for a large aluminium pool on the cell bottom to provide electrical contact for the cathodes. This will render the cell susceptible to the influence of the magnetic fields created by the busbar system, and may hence cause local short-circuiting of the electrodes.
Additionally, the referred patents, as well as U.S. Pat. No. 6,030,518, all point to the lowering of the bath temperature as compared to normal Hall-Hèroult cell temperatures as a means of a feasible reduction of the anode corrosion rates in the cell. The utilisation of the gas-lift effect and design of so-called up-corner and down-corner flow funnels are also described in U.S. Pat. No. 4,308,116, specially aimed at magnesium production.
U.S. Pat. No. 4,681,671 describes a novel cell design with a horizontal cathode and several, blade-shaped vertical anodes, and the cell is then operated at low electrolyte temperatures and with an anodic current density at or below a critical threshold value at which oxide-containing anions are discharged preferentially to fluoride anions. By means of forced or natural convection, the melt is circulated to a separate chamber or a separate unit, in which alumina is added before the melt is circulated back into the electrolysis compartment. Although the total surface area of the anode is high in the proposed configuration, the effective anode area is small and limited due to the low electrical conductivity of the anode material relative to the electrolyte. This will substantially limit the useful anodic surface area, and will lead to high corrosion rates at the effective anode surface.
A fact well established in hydrodynamics is that the flow of a fluid system is governed by a balance between the driving force for fluid flow and the resistance to fluid flow within the components of the system. Furthermore, depending upon the configuration, the velocity within local regions flow may be in the same direction but may sometimes be in the direction opposite to the fluid drive. This principle is amongst others cited in U.S. Pat. Nos. 3,755,099, 4,151,061 and 4,308,116. Inclined electrode surfaces are used to enhance/facilitate the drainage of gas bubbles from the anode and molten metal from the cathode. Hence, the design of electrolysis cells with vertical or near horizontal electrodes of both multi-monopolar and bipolar electrode arrangement, where fixed interpolar distance and the gas-lift effect are used to create a forced convection of the electrolyte flow, is not new. WO 02/31225 and U.S. Pat. Nos. 3,666,654, 3,779,699, 4,151,061 and 4,308,116, amongst others utilise such design principles, and the two latter patents also give descriptions of the use of “funnels” for up-comer(s) and down-comer(s) with respect to the electrolyte flow. U.S. Pat. No. 4,308,116 also suggests the use of a separation wall for enhanced separation of produced metal and gas. However, the inclined rod-shaped anodes described in WO 02/31225 do not set up Such a strong and controlled bubble driven flow as the present invention, and experiments show that gas will escape from all sides of such an anodes even if the bottom surface is inclined several degrees.